Real-time subtitles for everything you say and hear on Windows

What it does Ecoute is a Python tool that listens to your default microphone and speakers on Windows, then prints live transcripts of both streams into a single text box. It is built around OpenAI’s Whisper and is meant to help you keep track of conversations as they happen.

The interesting bit The tool captures both your own voice and the audio coming out of your speakers, effectively transcribing both sides of a call or meeting without needing access to the meeting software itself. That makes it a universal, if slightly brute-force, captioning layer for any Windows audio.

Key highlights

• Runs locally using Whisper tiny for English transcription without an internet connection (after setup).
• Optional --api mode routes audio to OpenAI’s Whisper API for faster, more accurate, multilingual transcription at the cost of API credits.
• Captures both microphone input and speaker output simultaneously, labeling them “You” and “Speaker”.
• Windows-only; the README explicitly notes it has not been tested on other operating systems.

Caveats

• Without the API flag, the local model is locked to English and may struggle with accents or uncommon words.
• It only listens to the default microphone and default speaker; switching audio devices requires changing Windows system defaults.
• The README warns that the --api flag is expected to become the default, which would make the offline mode secondary and tie core functionality to paid API access.

Verdict Handy for Windows users who want a quick, drop-in captioning layer for any call or in-person conversation. Skip it if you are on macOS or Linux, or if you need reliable transcription of non-English speech without paying for API tokens.